Expression of the liver 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase mRNA in FAO-1 cellsCarme ESPINET,* Alberto M. VARGAS,t M. Raafat EL-MAGHRABI, Alex J. LANGE and Simon J. PILKISDepartment of Physiology and Biophysics, Health Science Center, SUNY at Stony Brook, Stony Brook, NY 11794, U.S.A.

The hormonal regulation of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase gene expression was studied in the rat hepa-toma cell line FAO-i. Both 6-phosphofructo-2-kinase andfructose-2,6-bisphosphatase activities were detected in FAO-icells, at 680% of the levels found in rat liver. Northern blotanalysis showed that FAO-i cells, like rat liver, contained a

predominant species of bifunctional enzyme mRNA, which is2.2 kb in size. A sensitive RNAase protection assay revealed thepresence in FAO-i cells of an additional mRNA species, whichis generated when transcription is initiated from the skeletalmuscle promoter of the rat liver/skeletal muscle gene. Theliver/skeletal muscle mRNA ratio in FAO-I cells was 10:1,

INTRODUCTION

The functional specialization of the mammalian liver is dictatedby specific regulation of gene transcription in the terminallydifferentiated hepatocyte [1]. Studies with cultured hepatoma celllines have revealed that expression of many of the genes ofhepatic carbohydrate metabolism is regulated in a complexmanner by hormones such as glucagon, via its second messenger

cyclicAMP (cAMP), glucocorticoids and insulin [2,3]. The effectsof hormones on gene expression in hepatoma cells or primaryhepatocytes are, in general, consonant with their establishedphysiological actions: insulin increases the abundance ofmRNAsthat encode glycolytic enzymes [e.g. glucokinase, pyruvate kinase,6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (6PF-2-K/Fru-2,6-P2ase)] and decreases mRNAs that encode gluco-neogenic enzymes [e.g. phosphoenolpyruvate carboxykinase(PEPCK) and fructose-1,6-bisphosphatase (Fru-1,6-P2ase)];cAMP has the opposite effects. Glucocorticoid hormones, on theother hand, have a more limited action, playing an importantrole in increasing PEPCK and 6PF-2-K/Fru-2,6-P2ase mRNAabundances, and they also have a permissive action in theregulation of pyruvate kinase mRNA by insulin [2,3]. However,while the cis/trans model of gene regulation has been successfullyapplied to an analysis of the PEPCK gene promoter in hepatomacells, this approach has yielded only limited success with pro-moters for other genes, such as those for pyruvate kinase,glucokinase and 6PF-2-K/Fru-2,6-P2ase [2,3]. The fact that manyof these genes are not expressed or regulated in standard cellculture lines has hampered this type of analysis. For example,hormonal and carbohydrate control of the liver-type pyruvatekinase and glucokinase genes, that is readily demonstrated in vivo(see [2,3] for a review), does not occur in hepatoma cells [4], and

which is similar to that observed in rat liver. In contrast, inanother rat hepatoma cell line, FTO-2B, only the skeletal musclemRNA was detected. Insulin and dexamethasone induced theliver bifunctional enzyme mRNA in FAO- I cells by 2-4-fold and10-20-fold respectively in a concentration- and time-dependentmanner, and their effects were antagonized by cyclic AMP.Transcription of the gene in FAO-I cells, measured by nuclearrun-on assays, was also enhanced by dexamethasone and insulin.It is concluded that the FAO-I cell line is similar to liver withrespect to both the preferential use of the liver promoter of thegene and its regulation by hormones, and is therefore an excellentmodel for the study of the hepatic expression of this gene.

can only be studied in primary hepatocytes [5]. However, withthe exception of PEPCK, none of the liver-specific forms ofglycolytic and gluconeogenic enzymes are expressed in primaryhepatocyte cultures unless insulin and/or dexamethasone are

added to the media. Furthermore, primary hepatocyte culturespresent many technical problems, including reproducibility andsurvival time.The bifunctional enzyme 6PF-2-K/Fru-2,6-P2ase catalyses the

synthesis and hydrolysis of fructose 2,6-bisphosphate (Fru-2,6-P2), a potent modulator of both glycolysis and gluconeo-genesis, through its action on the activities of 6-phosphofructo- 1-

kinase (6PF-l-K) and Fru-1 ,6-P2ase [6,7]. 6PF-2-K/Fru-2,6-P2ase is expressed in all mammalian tissues [8,9]. The liver andskeletal muscle isoforms are different with respect to their N-terminal sequences, phosphorylation/dephosphorylation reac-

tions and patterns of regulation by metabolites [7-9]. Bothisoforms are products of a single 55 kb gene which contains 15exons [10,11], with the muscle-specific transcript being initiatedjust distal to an upstream promoter containing exon 1M and fromwhich the liver exon 1L is spliced out during processing. Theliver-specific transcript is initiated just distal to a second pro-

moter, upstream of exon 1L, and generates exon 1L. Either ofthese first exons (1L or 1M) is then spliced to the common exons

2-14. Cell-specific regulation of the liver and skeletal muscleforms of the enzyme is thus accomplished by the alternative firstexons providing unique mRNAs in these tissues [2,3].

Both insulin and dexamethasone enhance bifunctional enzymegene transcription in the rat hepatc.ma cell line FTO-2B [12],although the mRNA expressed was shown to be related to theskeletal muscle transcript. This finding raises doubt as to whetherthe same hormone response elements in the gene are involved inthe regulation of the liver-specific mRNA that has been observed

Abbreviations used: cAMP, cyclic AMP; PEPCK, phosphoenolpyruvate carboxykinase; 6PF-1-K, 6-phosphofructo-1-kinase; 6PF-2-K/Fru-2,6-P2ase,6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase; Fru-1,6-P2ase, fructose-1,6-bisphosphatase; Fru-2,6-P2, fructose 2,6-bisphosphate; Fru-6-P,fructose 6-phosphate.

* Present address: Department Bioquimica, Facultat de Medicina, Estudi General de Lleida, Lleida, Spain.t Present address: Departamento Bioquimica y Biologia Molecular, Facultad de Ciencias, 18071 Granada, Spain

173Biochem. J. (1993) 293, 173-179 (Printed in Great Britain)

174 C. Espinet and others

in vivo [2,3]. In order to develop a cell culture model moresuitable for studying regulation of the liver mRNA, we screeneda number of rat hepatoma cell lines in an attempt to find a linethat has a pattern of 6PF-2-K/Fru-2,6-P2ase gene expression andhormonal regulation similar to that found in rat liver in vivo. Inthis paper we report that FAO-i cells meet these criteria.

MATERIALS AND METHODS

MaterialsThe random-primed DNA labelling kit, restriction endo-nucleases, nucleotides, bovine insulin and dibutyryl cAMP werepurchased from Boehringer Mannheim. [a-32P]dATP, [a-35S]dATP, [a_-32P]UTP and [3H]uridine were from ICN Bio-medicals. GeneScreen membrane was from Du Pont-NewEngland Nuclear. Bicinchoninic acid (BCA) protein assay reagentwas obtained from Pierce Chemical Corp. Fetal bovine serumand cell culture media were from Gibco. Other materials andchemicals were of the highest quality available.

Cell cultureMonolayer cultures of rat FAO-1 hepatoma-derived cells weregrown in RPMI 1640 medium containing 10 mM glucose andsupplemented with 10% fetal bovine serum. FTO-2B cells weremaintained in culture as described previously [12]. All experi-ments were carried out before the cell density reached confluency

and cells were incubated in serum-free medium during thehormonal treatment. Hormone effects were tested in the absenceof serum to ensure that trace hormone concentrations in serumwould not confound the results. The results are presented relativeto untreated controls, incubated in the absence ofserum, analysedat the same time as the treated cells. Since the cells were culturedin serum until confluency and then without serum for the durationof the experiment, it is possible that the basal level of expressionchanged during the experiment. In order to test whether theabsence of serum in itself had any effect, cells were incubatedwith and without serum for up to 12 h. Relative to cells incubatedwith serum, the relative mRNA abundance (1.0 at zero time) was0.90+ 0.09 after 3 h, 0.95+ 0.07 after 6 h, 1.30 + 0.20 after 9 hand 1.20 + 0.15 after 12 h (means+ S.E.M. from three experi-ments). The results indicate that removal of serum had little orno effect during the 12 h time course.

Measurement of 6PF-2-K/Fru-2,6-P2ase activityFAO-1 cells (- 4 x 107) were homogenized in medium composedof 100 mM KC1, 20 mM Tes, pH 7.8, 5 mM potassiumphosphate, 5 mM EDTA, 5 mM EGTA, 1 mM dithiothreitol,0.5 mM phenylmethanesulphonyl fluoride and 2.5 mg/l leu-peptin. The enzyme was partially purified by a 6-21 % poly-(ethylene glycol) 8000 precipitation [13]. Under these conditionsthe in situ phosphorylation state of the enzyme is preserved [13].6PF-2-K activity was assayed by measuring the rate ofFru-2,6-P2 production from fructose 6-phosphate (Fru-6-P) andMgATP at pH 7.8 [13,14]. Fru-2,6-P2 was assayed by its abilityto activate pyrophosphate-dependent 6PF-1-K from potatotubers (Sigma) [15]. Fru-2,6-P2ase activity was assayed bymeasuring [32P]P1 release from [2-32P]Fru-2,6-P2 (10 ,uM) [14] orby phosphoenzyme formation [14,16].

(a)I

- -- .I Bil I1. .1

I I laI I I VT1 .1II 1 l2 3456 7 8 9 10 11 121314 DNA hybridization probes

probes220 bp 299 bp

F F R B R

Exon 1M Exon 1L

Muscle-specificprobe(C)

Liver-specificprobe(B)

(b)

C

Four different hybridization probes were used to detect 6PF-2-0 5 10 kb K/Fru-2,6-P2ase mRNA (Figure la). A 1.4kb EcoRI fragment

was isolated from the cDNA for 6PF-2-K/Fru-2,6-P2ase that1.4 kb contained sequences transcribed from the third to the fourteenth

exons, which are common to 6PF-2-K/Fru-2,6-P2ase from bothR liver and skeletal muscle [10,17,18]. A second probe was a 0.5 kb

14

EcoRI fragment containing the 5' end of the cDNA for theExons 3-14 0.5 kb bifunctional enzyme [13]. In order to detect tissue-specific mRNA,ommon probe expression, DNA fragments containing either the liver- or the

muscle-specific first exon were isolated from a genomic clone(A) (Figure la). The liver-specific probe is an EcoRI-BanII fragment

(299 bp) that includes most of exon 1,L while the muscle-specificprobe is a FokI (220 bp) fragment that contains exon IM. Theabundance ofPEPCK mRNA was determined with a 1.5 kb PstI

22.2 kb fragment isolated from a cDNA clone (pPCKlOrc) for the*-2.2 kb enzyme [19]. All DNA probes were random-primer labelled with

[OC-32P]dATP [20].

1 2 3

Figure 1 Skeletal muscle/liver biuncUonal enzyme gene structure andhybridization probes for Northern blot analysis

(a) Structure of the rat 6PF-2-K/Fru-2,6-P2ase liver/skeletal muscle gene. Solid bars representexons and connecting lines represent introns. The sequences used as hybridization probes arealso indicated below the structure of the gene. B, Banll; F, Fokl; R, EcoRl. (b) Northern blotof FAO-1 cell RNA (20 ,ug) (lane 1), rat liver RNA (10 ,ug) (lane 2) and rat skeletal muscle RNA(20 ,g) (lane 3). Hybridization was performed with the liver-specific DNA probe (probe B)shown in (a) as described in the Materials and methods section.

RNA Isolation and analysisRNA was isolated essentially according to Chirgwin et al. [21],with one extraction in 4 M guanidine thiocyanate followed bytwo successive extractions with 7.5 M guanidine hydrochloride.The abundance of the mRNA for 6-PF-2-K/Fru-2,6-P2ase was

determined either by Northern blot analysis [22] or by RNAaseprotection assay [23]. In both cases densitometric scans of theautoradiograms were carried out on an LKB-Pharmacia Ultro-Scan XL Laser Densitometer.

I I1M 1

cDNA

.mp

6-Phosphofructo-2-kinase/fructose-2,6-bisphosphatase gene regulation

Ribonuclease protection assayAn RNA probe was constructed according to Maniatis et al. [23].The 0.5 kb EcoRI fragment, previously described [13], was clonedinto the EcoRI restriction site of the plasmid Bluescript SK(+).The plasmid was linearized by digestion with Avall, and wasthen used to synthesize RNA using its T3 RNA polymerasepromoter, with ribonucleotide triphosphates, including [a-32P]-CTP, and T3 RNA polymerase. The synthesized riboprobewas 390 bases long, complementary to 319 bases of the liver-specific mRNA, including exon 1,L and to 192 bases of theskeletal muscle-specific mRNA. The probe was purified byelectrophoresis in 60% polyacrylamide, in the presence of 8 Murea and Tris/borate/EDTA buffer. Samples of 10-50,g oftotal RNA from extracts of nearly confluent cells, and from ratliver and rat muscle, were incubated at 90 °C for 10 min with amolar excess (100000 c.p.m.) of the purified riboprobe. Thetubes were cooled slowly to 42 °C, and maintained at thistemperature overnight. The assay was performed as described byHod [24]. A mixture of RNAase TI and RNAase A were used todegrade single-stranded unhybridized probe. After digestion, theprotected fragments were collected by centrifugation (12000 g;15 min) at 4 °C in the presence of guanidine thiocyanate andpropan-2-ol, and subjected to electrophoresis on DNAsequencing gels, as described previously [24]. The gels were thensubjected to autoradiography and densitometric scanning.

Measurement of run-on gene transcription in Isolated nucleiThe isolation of nuclei and the incorporation of [a-32P]UTP intonascent RNA transcripts of 6PF-2-K/Fru-2,6-P2ase were per-formed as previously described [12].

cDNA library construction"and screeningA cDNA library was constructed by using mRNA isolated fromFAO-I cell extracts, a cDNA synthesis kit from Pharmacia, anda A ZAP II cloning kit from Stratagene. In order to obtain the 5'region of the 6PF-2-K/Fru-2,6-P2ase gene, four different specificoligonucleotides, located in the common region of the gene, wereused to prime the reverse transcriptase reaction. The library wasscreened with the 0.5 kb EcoRI cDNA probe [13]. Positive cloneswere excised in vivo with R408 helper phage.

DNA sequencing and analysisAll sequencing was done by the Sanger dideoxy method [25]using the Sequenase sequencing kit (U.S. Biochemicals, SanDiego, CA, U.S.A.). Sequence analysis was carried out using theGCG software package [26].

Protein determinationProtein concentration was determined with the BCA ProteinAssay Reagent from Pierce Chemical Co., using the enhancedprotocol and with BSA as a standard.

RESULTSExpression of 6PF-2-KIFru-2,6-P2ase In FAO-1 cellsExtracts from FAO-i cells were assayed for 6PF-2-K activity atpH 7.8 with saturating concentrations of substrates (Fru-6-P,

5 mM; MgATP, 5 mM). Under these assay conditions, Vmax.values reflect the amount of enzyme protein [13]. FAO-i cellscontained about 27 ,4units of 6PF-2-K activity per mg of cellularprotein, while cells exposed for 12 h to insulin (0.1 ,uM) ordexamethasone (1 ,uM) had 6PF-2-K activities of about 61 and81 ,tunits/mg respectively. The increase in 6PF-2-K activity wasseen after a lag of 4-6 h. The extract from FAO-I cells alsoexhibited Fru-2,6-PJase activity (20 uunits/mg ofprotein) (resultsnot shown). In addition, we identified, by SDS/PAGE, a 32p_labelled polypeptide in extracts incubated with [2-32P]Fru-2,6-P2that had a molecular mass of 55 kDa, which is identical to thesize of the 6PF-2-K/Fru-2,6-P2ase subunit from rat liver (resultsnot shown). As with 6PF-2-K activity, a quantitatively similarenhancement in Fru-2,6-P2ase activity was observed in cellstreated with insulin or dexamethasone. The activity ratio ofkinase to bisphosphatase was about 1.4, which is similar to thatfor the liver enzyme, but different from the activity ratios of0.1-0.5 and 50 found in skeletal muscle and heart respectively [8].The 6PF-2-K activity in livers from normal fed rats is approx.40,units/mg of cellular protein [13]. The presence of 6PF-2-K/Fru-2,6-P2ase in FAO-i rat hepatoma cells was furthercorroborated by Western blot analysis of the cellular extractusing an antibody against rat liver 6PF-2-K/Fru-2,6-P2ase,which also revealed a 55 kDa subunit. These results indicate thatthe enzyme in FAO-I cells is similar in amount and properties to6PF-2-K/Fru-2,6-P2ase from rat liver.

Identfflcation of the mRNAs for 6PF-2-K/Fru-2,6-P2ase in FAO-1cellsThe 5' region of the skeletal muscle mRNA contains sequencesderived from an exon located about 5 kb upstream of the liver-specific exon (see Figure la for the structure of the 6PF-2-K/Fru-2,6-P2ase gene). The remaining 13 exons are common toboth forms of mRNA. A 1.4 kb cDNA probe [13,27] containingsequences common to liver and muscle mRNA (exons 3-14)hybridized with a 2.2 kb mRNA in FAO-I cells (results notshown). A liver-specific probe hybridized with a 2.2 kb RNAfrom both FAO-i cells and liver, but not with muscle RNA(Figure lb). When the muscle-specific probe was used, a signalwas observed only in the muscle RNA lane (results not shown),suggesting that the liver 6PF-2-K/Fru-2,6-P2ase mRNA is thepredominant species expressed in FAO-I cells. It is not possible,however, to rule out the existence of other low-abundance 6PF-2-K/Fru-2,6-P2ase mRNAs the concentrations of which may bebelow the detection limit of Northern blot analysis, or whose sizeand sequence are similar to those of the liver mRNA. Liver andskeletal muscle mRNAs differ in size by less than 200 bases, andcannot be easily distinguished by Northern blot analysis.More conclusive results were obtained by an RNAase pro-

tection assay (Figure 2). Use of the 390-base RNA probepermitted a clear discrimination between the liver and musclemRNAs (Figure 2). As predicted, the major protected fragmentin liver RNA was 319 bases long, corresponding to the regioncomplementary to the liver mRNA, while the major protectedfragment in skeletal muscle RNA was 192 bases long. Noprotected bands were observed with yeast RNA. While liver andskeletal muscle contained both the liver and skeletal muscletranscripts, their abundances are clearly different in the twotissues: in liver the liver transcript is the predominant form, whilein skeletal muscle the reverse is true. Quantification revealed thatthe liver mRNA constituted approx. 15% of the total bi-functional mRNA in skeletal muscle, while the skeletal musclemRNA accounted for only about 10% of the total bifunctionalmRNA in liver (Table 1). This result shows that both mRNAs

175

176 C. Espinet and others

(a)

Liver exon Common exons

IIIAvail | 1 -5' 3'.4 _...._

192 basesMuscle-protected fragment

319 basesLiver-protected fragment

(c)390- -

IbpI-319

Time (h)

-192

1 2 3 4 56

:B..::.. 2 4 -192

1 23 4

Figure 2 RNAase protectUon analysi(a) The structure of the 500 bp fragment of the rat liver 6PF-2-K/Fru-2,6-P2ase EcoRI cDNA aftersubcloning into Bluescript( +) and linearization by digestion with Avail is shown at the top. Theregions of the probe predicted to be protected by the skeletal muscle and liver RNA afterRNAase digestion are shown below the 390-base probe. (b) RNAase protection assay of RNAfrom FAQ-i cells, rat liver and rat skeletal muscle. Lane 1, undigested RNA probe (390 bases);the other lanes contain protected fragments after hybridization of the probe with RNA fromdifferent sources and digestion with RNAase: yeast RNA (20 ,u.g) (lane 2), skeletal muscle(15 1ug) (lane 3), rat liver (5 ,ug) (lane 4), FAQ-l cells (50 and 20,lg) (lanes 5 and 6respectively). (c) RNAase protection assay of RNA from liver (10 ,ug) (lane 2), FT0-2B cells(50 ,sg) (lane 3) and skeletal muscle (20 ug) (lane 4). The undigested RNA probe is shownin lane 1.

Table 1 Abundance of the skeletal muscle and liver 6PF-2-K/Fru-2,6-P.asemRNAs in rat hepatoma cells relative to the abundance of the liver mRNAIn liverSkeletal muscle and liver mRNA abundances were assayed by RNAase protection as describedin the Materials and methods section. The data are expressed relative to the abundance of theliver-specific mRNA (319 base protected fragment) in livers from fed rats, which was given avalue of 1.0. The mRNAs were quantified by scanning laser densitometry of the 319 and 192base protected fragments. The experiments represent the means + S.E.M. of three experiments.

Relative mRNA abundance

Tissue/cells Liver mRNA Skeletal muscle mRNA

Liver from fed ratsSkeletal muscle !.;om fed ratsFAQ-i cellsFTO-2B cells

1.00+ 0.110.15+ 0.010.19+ 0.020

0.10+ 0.010.91 +0.080.02 + 0.010.14 + 0.02

are expressed in these tissues and suggests that the use ofalternate promoters of this gene is tissue-specific.

It was of interest to ascertain whether both bifunctionalenzyme mRNAs are also expressed in hepatoma cell lines derived

Figure 3 Time courses of hormonal effects on SPF-2-K/Fru-2,6-P2asemRNA Induction

Cells were incubated in medium containing either 1 #uM insulin (O) or 1 ,uM dexa-methasone (El). Fetal serum was washed out from the incubation plate at time 0, just beforethe addition of the hormones. 6PF-2-K/Fru-2,6-P2ase mRNA was determined by Northern blotanalysis, and the changes in mRNA abundance are presented relative to untreated controls,incubated in the absence of serum, analysed at the same time as the treated cells, Dex,dexamethasone; Ins, insulin.

from rat liver. Application of the RNAase protection assay toRNA from FAO-i cells showed a pattern of tissue-specificexpression that is identical to that of rat liver (Figure 2b; Table1), albeit at one-fifth the abundance. This is in contrast toenzyme activity levels, which are about two-thirds of those inliver. The explanation for this discrepancy is not clear, but mayrelate to different rates of translation and/or protein degradationin the FAO-I cells versus the liver. Figure 2(c) shows that onlythe skeletal muscle mRNA is expressed in another rat hepatomacell line, FTO-2B, a finding consistent with earlier studies ofFTO-2B cell 6PF-2-K/Fru-2,6-P2ase mRNA, which was recog-nized by a skeletal muscle-specific cDNA probe but not by aliver-specific probe [12]. Table 1 shows that the abundance ofskeletal muscle mRNA in FTO-2B cells is slightly greater thanthat of the skeletal muscle mRNA in rat liver.

In order to rule out a possible dissimilarity in the sequences ofthe 5' ends of the liver and FAO-I cell mRNAs, an FAO-IcDNA library was screened as described in the Materials andmethods section, and a clone containing the 5' end of the 6PF-2-K/Fru-2,6-P2ase gene was isolated. Its sequence was identicalto that of the first 328 bases of the liver 6PF-2-K/Fru-2,6-P2asecDNA (results not shown). This finding supports the hypothesisthat the predominant FAO- I cell bifunctional enzyme mRNA isidentical to the liver-specific mRNA.

Hormonal control of 6PF-2-K/Fru-2,6-P2ase gene expression inFAO cellsFigure 3 shows the time course of 6PF-2-K/Fru-2,6-P2ase livermRNA accumulation in FAO-i cells incubated in the presenceof 1 ,uM insulin or 1 ,uM dexamethasone. In both insulin- anddexamethasone-treated cells the abundance of the liver mRNAincreased gradually, and attained maximum levels that were 2-4-fold and 20-fold higher respectively than the basal level. Theincrease in the level of 6PF-2-K/Fru-2,6-P2ase mRNA in insulin-treated cells was noted only after a 4-6 h lag, while the onset

T3 promoter

390 base probe

(b)

c-.:

22

20

18

16

14

12

10

8

6

4

2

z

E0

(0

C4

0.(0(bp)

-319

6-Phosphofructo-2-kinase/tructose-2,6-bisphosphatase gene regulation

(a)

ml .4Lr - 2.2 kb

(bi

_ W* O_011 22.8kb

1 l 3 4 5 6 7 8

Figure 4 Hormonal regulation of 6PF-2-K/Fru-2,6-P2ase mRNA abundancein FAO-1 cells

Northern blot analysis of RNA (20 1ug) extracted from cells incubated for 12 h with no additions(lane 1), 0.1 1uM insulin (lane 2), 1 ,sM insulin (lane 3), 0.1 ,uM dexamethasone (lane 4), 1 #Mdexamethasone (lane 5), 10 ,uM dexamethasone (lane 6), 0.5 mM dibutyryl cAMP plus 1 mMtheophylline (lane 7) or 0.5 mM monobutyryl cAMP plus 1 mM theophylline (lane 8). In (a),a 1.4 kb EcoRl fragment (common probe; see Figure la) from the 6PF-2-K/Fru-2,6-P2ase ratliver cDNA was used as a probe. In (b), the probe was a 1.5 kb Pstl fragment from the cDNAfor PEPCK.

0 10- 10-7Concn. (M)

10-6 10

Figure 5 Dose-response of insulin and dexamethasone action of 6PF-2-K/Fru-2,6-P2ase mRNA abundance in FAO-l cells

FAO cells were incubated in media containing the indicated concentrations of insulin (Ins; *)and dexamethasone (Dex; [1) for 12 h. The abundance of 6PF-2-K/Fru-2,6-P2ase mRNA wasdetermined by Northern blot analysis using the 1.4 kb common probe, and quantified byscanning densitometry. The results are expressed as described in the Materials and Methodssection and in the legend to Figure 3.

of the response to dexamethasone was seen as early as 3 h,suggesting that the mechanism of induction of 6PF-2-K/Fru-2,6-P,ase mRNA may differ for the two hormones.As can be seen in Figure 4(a), 0.1 ,uM insulin was sufficient to

increase the 6PF-2-K/Fru-2,6-P2ase mRNA abundance (lane 2),although a much larger increase was observed in the presence ofdexamethasone (lanes 4, 5 and 6). As expected, the content ofPEPCK mRNA in FAO-I cells was increased by dexamethasoneand cAMP (Figure 4b, lanes 4-6, and 7 and 8, respectively).

'a

-az

E0

C4tD

*,

LL0-(0

10

InsulinDexMb-cAMPDb-cAMPTheo

+ +

- + + - + + -_ _ + + + _-

+ + +

+ +

Figure 6 Effect of cAMP on the induction by Insulin and dexamethasone of6PF-2-1/Fru-2,6-P2ase mRNA In FAO-l cells

The cells were incubated for 12 h under different conditions. Final concentrations were 1 #Mfor insulin and dexamethasone (Dex), 0.5 mM for cAMP analogues (Mb-cAMP, monobutyrylcAMP; Db-cAMP, dibutyryl cAMP) and 1 mM for theophylline (Theo). The RNA was analysedas in Figure 4, and the results are expressed as described in the Materials and methods sectionand the legends to Figures 3 and 5. The experiment was repeated three times with essentiallysimilar results.

Dibutyryl cAMP plus theophylline, in the absence of insulin or

dexamethasone, had no effect on 6PF-2-K/Fru-2,6-P2asemRNA abundance.

Figure 5 shows dose-response curves of the induction of 6PF-2-K/Fru-2,6-P2ase liver mRNA in FAO-I cells exposed to insulinor dexamethasone. Half-maximal effects of insulin and dexa-methasone were obtained in the presence of 0.01 #uM of eitherhormone (Figure 5), concentrations that are within the physio-logical range for these hormones.

Dibutyryl cAMP Inhibits the induction of 6PF-2-K/Fru-2,6-P2asemRNA by insulin and dexamethasoneInduction of 6PF-2-K/Fru-2,6-P2ase mRNA by either insulin or

dexamethasone was counteracted by the simultaneous additionof cAMP analogues to the incubation medium (Figure 6).Monobutyryl cAMP was sufficient to counteract the insulininduction of the gene, whereas the presence of both theophyllineand dibutyryl cAMP were required to antagonize the much more

potent effect of dexamethasone. The cAMP analogues repressedthe effect of saturating concentrations of dexamethasone andinsulin, suggesting that they play a dominant role in regulatingthis gene's expression in FAO-I cells. A similar dominant role ofcAMP in affecting the induction of the skeletal muscle bi-functional enzyme mRNA has been reported for FTO-2B cells[12].

Induction of 6PF-2-K/Fru-2,6-P2ase mRNA is mediated byactivation of gene transcriptionThe increase in mRNA may be explained by an increase in itsrate of synthesis, by a decrease in its degradation rate, or by a

combination of both mechanisms. Run-on transcription assayswith isolated nuclei were performed to estimate the relativetranscription rate of 6PF-2-K/Fru-2,6-P2ase gene in FAO-icells. As shown in Figure 7(a), there was an increase in the

177

10

8 -

6-

4-

2 Iii

n I I I I I

.3

z

En

CoC4

LL.a.Co

178 C. Espinet and others

10(a) (b)

= 8 6PF-2-K/Fru-2,6-P2ase PEPCK

P6

.2 4

2

0

[ Control Insulin L Dexamethasone

Figure 7 Effect of insulin and dexamethasone on 6PF-2-K/Fru-2,6-P2asegene transcripfton In isolated nuclei

Nuclei were isolated from cells incubated for 9 h in the presence of either 1 ,uM insulin or 1 ,#Mdexamethasone. The nuclei were used in run-on transcription assays which were carried outas described in the Materials and methods section. Means+ S.E.M. from three independentexperiments are shown.

relative transcription rate of the gene in cells exposed to eitherinsulin or dexamethasone for 9 h. The time courses of inductionof 6PF-2-K/Fru-2,6-PJase gene transcription and mRNA abun-dance (Figure 3) showed that the rate of mRNA accumulationgenerally followed the changes in the relative synthesis rate of themRNA (results not shown). These results strongly suggest thatthe mechanism for the increase in the 6PF-2-K/Fru-2,6-P2asemRNA level in cells exposed to insulin or dexamethasone ismediated, in large part by stimulation of the transcription rate ofthe gene. For comparison, the transcriptional activity of the genecoding for PEPCK was also analysed in the same experiment(Figure 7b). As expected, the transcription of the PEPCK genewas enhanced in cells treated with dexamethasone and inhibitedin cells incubated in the presence of insulin. The relatively smalleffects of insulin and dexamethasone on PEPCK transcriptionrelate to the fact that measurements were done after 9 h. Asimilar extent of inhibition by insulin of PEPCK gene tran-scription has been documented previously [28]. The differentialresponse in the relative transcription rates of the 6PF-2-K/Fru-2,6-P2ase and PEPCK genes in response to insulin and dexa-methasone indicates that the enhancement in the transcriptionrate of 6PF-2-K/Fru-2,6-P2ase gene is specific.

DISCUSSIONStudies on the control of skeletal muscle/liver 6PF-2-K/Fru-2,6-P2ase gene expression have been done in vivo and in rathepatoma FTO-2B cells. The amount of bifunctional enzymeprotein decreases in livers of rats during starvation or diabetes,and is restored by refeeding a high-carbohydrate diet or byinsulin administration respectively [13,29]. The amount of 6PF-2-K/Fru-2,6-P2ase mRNA is also reduced in liver and skeletalmuscle of adrenalectomized rats, and is restored by administra-tion of glucocorticoids which increase gene transcription[11,17,27]. This is a direct effect of glucocorticoids, since theiraddition to primary cultures of hepatocytes rapidly increasesmRNA levels by nearly 100-fold [17]. Dexamethasone and insulin

mRNA, not a liver mRNA ([12]; the present paper). This paper

clearly shows that, in contrast to FTO-2B cells, FAO-i cellsprovide an excellent model system for studying the hormonalregulation of the liver-specific bifunctional enzyme mRNA.The results of the present study also demonstrate that both the

liver and skeletal muscle forms are expressed in muscle, liver andFAO-I cells. The glucokinase gene provides another example oftissue-specific expression of different isoforms as a result of use

of alternative first exons to generate pancreatic ,f-cell and liverisoforms [2]. However, the weight of evidence indicates that thefl-cell glucokinase mRNAs are expressed only in f-cells and theliver glucokinase mRNAs are expressed only in liver, whichsuggests that, in contrast to the skeletal muscle/liver bifunctionalenzyme gene, alternative promoter use in the glucokinase gene isstrictly tissue-specific [30].The use of both the muscle and liver promoters of the 6PF-2-

K/Fru-2,6-P2ase gene in liver and skeletal muscle raises thequestion of whether the effects of hormones on bifunctionalenzyme gene expression are differentially regulated, i.e. dohormones increase liver promoter activity in liver but not skeletalmuscle promoter activity in muscle? Although data on thisquestion are still limited, preliminary results with the RNAaseprotection assay (A. Vargas, unpublished work) have shown thatboth mRNAs were increased by insulin and glucocorticoids inFAO-i cells, and previous work has demonstrated the sameresult in both skeletal muscle and liver of adrenalectomized ratsafter glucocorticoid administration [27]. Furthermore, althoughonly the skeletal muscle transcript is expressed in FTO-2B cells,regulation of this transcript by insulin, dexamethasone andcAMP is similar to that observed for the liver-specific transcriptin vivo and in primary cultures of hepatocytes [31], except thatdexamethasone is a more potent inducer of 6PF-2-K/Fru-2,6-P2ase gene expression than insulin in vivo [13,29], in primaryhepatocyte cultures [31] and in FAO-i cells (present paper), butthe reverse is true in FTO-2B cells [12]. The reason for thisdifference in hormone responsiveness is not known.

Analysis of RNA isolated from rat hepatoma H35 cells, fromwhich both FTO-2B and FAO-i cells were derived [32], alsorevealed a pattern of bifunctional enzyme gene expression that issimilar to those of FAO-i cells and rat liver (A. Vargas,unpublished work). Since FTO-2B cells were originally generatedby treating a derivative cell line of H35 cells with mutagens andvisible light in order to select for thymidine kinase-deficient cells[33], it is possible that this treatment and/or subsequent selectivepressures on the cell line have resulted in an apparent loss ofliver-specific factors that activate the liver promoter.The finding that bifunctional enzyme gene expression is subject

to regulation by the same set of hormones in FAO-I and FTO-2B cells suggests that the same hormone response elements act onboth promoters. Consistent with this idea, a unique gluco-corticoid response unit, recently identified in the first intron ofthe gene, has been shown to enhance the activity of bothpromoters in transiently transfected cells [II]. This finding, andthe results of LeMaigre et al. [34] and Darville et al. [35], arguestrongly for the existence of two separate functional promotersof the gene. Based on the results presented here, it is reasonableto postulate the existence of tissue-specific regulatory systemswhich determine the relative activities of the liver and skeletalmuscle promoters in their respective tissues. Previous work hasdemonstrated the presence of numerous binding sites for trans-acting factors in the 5'-flanking regions of the gene that influencethe activities of both the liver and skeletal muscle promoters[34,35]. However, tissue-specific transcriptional factors respon-

also induce bifunctional enzyme mRNA in FTO-2B cells, butthese cells express a skeletal muscle 6PF-2-K/Fru-2,6-P2ase

sible for the pattern of alternative promoter use in liver, FAO- Icells and FTO-2B' cells were not identified. Work is in progress,

6-Phosphofructo-2-kinase/fructose-2,6-bisphosphatase gene regulation

using the rat hepatoma FAO-I and FTO-2B cell lines as

convenient model systems, to define the basis for first-exonselection of the skeletal muscle/liver gene.

A.M.V. was partially supported by grants from Junta de Andalucia and Ministerio deEducation y Ciencia, Spain. This work was supported by NIH grant no. DK3835405to S.J.P.

REFERENCES1 Dermon, E., Krauter, K., Wailing, L., Weinberger, C., Ray, M. and Darnell, Y. (1981)

Cell 23, 731-7392 Granner, D. and Pilkis, S. J. (1990) J. Biol. Chem. 265, 10173-101763 Pilkis, S. J. and Granner, D. (1992) Annu. Review Physiol. 54, 885-9094 Meinenhofer, M. C., deMediers, E., Cognet, M. and Kahn, A. (1987) Eur. J. Biochem.

169, 237-2435 Decaux, J. F., Antione, B. and Kahn, A. (1989) J. Biol. Chem. 264, 11881-118906 Pilkis, S. J., El-Maghrabi, M. R. and Claus, T. H. (1988) Annu. Rev. Biochem. 57,

755-7837 Pilkis, S. J. (ed.) (1990) Fructose-2,6-Bisphosphate, CRC Press, Boca Raton, FL8 Hue, L., Rider, H. and Rousseau, G. G. (1990) in Fructose-2,6-Bisphosphate (Pilkis,

S. J., ed.), pp. 173-192, CRC Press, Boca Raton, FL9 Pilkis, S. J. and Claus, T. H. (1991) Annu. Rev. Nutr. 11, 465-515

10 Darville, M. I., Crepin, K. M., Hue, L. and Rousseau, G. G. (1989) Proc. Natl. Acad.Sci. U.S.A. 86, 6543-6547

11 Lange, A. J., Espinet, C., Hall, R., El-Maghrabi, M. R., Vargas, A. M., Miksicek, R. J.,Granner, D. K. and Pilkis, S. J. (1992) J. Biol Chem. 267, 15673-15680

12 Cifuentes, M., Espinet, C., Lange, A. J., Pilkis, S. J. and Hod, Y. (1991) J. Biol.Chem. 266, 1557-1563

13 Colosia, A. D., Marker, A. J., Lange, A. J., El-Maghrabi, M. R., Granner, D. K., Tauler,A., Pilkis, J. and Pilkis, S. J. (1988) J. Biol. Chem. 263, 18669-18677

14 El-Maghrabi, M. R., Correia, J. J., Heil, P. J., Pate, T., Cobb, C. and Pilkis, S. J.(1986) Proc. Natl. Acad. Sci. U.S.A. 83, 5005-5009

15 Van Schaftingen, E., Lederer, B., Bartrons, R. and Hers, H. G. (1982) Eur. J.Biochem. 129, 191-195

16 Stewart, H. B., El-Maghrabi, M. R. and Pilkis, S. J. (1985) J. Biol. Chem. 260,12935-12941

17 Lange, A. J., Kummel, L., El-Maghrabi, M. R., Tauler, A., Colosia, A. D., Marker, A.and Pilkis, S. J. (1989) Biochem. Biophys. Res. Commun. 162, 753-760

18 Crepin, K. M., Darville, M. I., Hue, L. and Rousseau, G. G. (1989) Eur. J. Biochem.183, 433-440

19 Yoo-Warren, H., Monahan, J. E., Short, H., Bruzel, A., Wynshaw-Boris, A., Meisner,H. M., Samols, D. and Hanson, R. W. (1983) Proc. Natl. Acad. Sci. U.S.A. 80,3656-3660

20 Feinberg, A. P. and Vogelstein, B. (1984) Anal. Biochem. 137, 266-26721 Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. and Rutter, W. J. (1979)

Biochemistry 18, 5294-529922 Hod, Y., Morris, S. M. and Hanson, R. W. (1984) J. Biol. Chem. 259, 15603-1560823 Maniatis, T., Fritsch, F. and Sambrook, J. (1982) Molecular Cloning. A Laboratory

Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY24 Hod, Y. (1992) Biotechniques 13, 852-85325 Sanger, F., Nicklen, S. and Coulson, R. R. (1977) Proc. Natl. Acad. Sci. U.S.A. 74,

5463-546726 Devereaux, J., Haeberli, P. and Smithies, 0. (1984) Nucleic Acids Res. 10, 70327 Marker, A. J., Colosia, A. D., Tauler, A. D., Solomon, D., Cayre, Y., Lange, A. J., El-

Maghrabi, M. R. and Pilkis, A. J. (1989) J. Biol. Chem. 264, 7000-700428 Granner, D. K., Andreone, T., Sasaki, K. and Beale, E. (1983) Nature (London) 305,

549-55129 Crepin, K. M., Darville, M. L., Hue, H. L. and Rousseau, G. G. (1988) FEBS Lett.

227,136-13930 Magnuson, M. (1992) J. Cell. Biochem. 48, 115-12131 Kummel, L. and Pilkis, S. J. (1990) Biochem. Biophys. Res. Commun. 169, 406-41332 Rogier, E., Cassio, D., Weiss, M. and Feldman, G. (1986) Differentiation 30, 329-33633 Killary, A. M., Lugo, T. M. and Fournier, R. E. K. (1984) Biochem. Genet. 22,

201-21334 LeMaigre, F. P., Durrvax, S. M. and Rousseau, G. G. (1991) Mol. Cell. Biol. 11,

1099-110635 Darville, M. I., Antoine, I. V. and Rousseau, G. G. (1992) Nucleic Acids Res. 20,

3515-3583

Received 25 November 1992/2 February 1993; accepted 5 February 1993

179